xref: /llvm-project/llvm/test/CodeGen/AMDGPU/si-fold-copy-kills.mir (revision 1b560e6ab7fc16c915bb054d3bffe5f77e08c3d5)
1# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
2# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -run-pass=si-fold-operands -verify-machineinstrs -o - %s | FileCheck --check-prefix=GCN %s
3
4---
5name: fold_reg_kill
6tracksRegLiveness: true
7body: |
8  bb.0:
9    liveins: $sgpr0
10
11    ; GCN-LABEL: name: fold_reg_kill
12    ; GCN: liveins: $sgpr0
13    ; GCN-NEXT: {{  $}}
14    ; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
15    ; GCN-NEXT: [[S_ADD_U32_:%[0-9]+]]:sreg_32 = S_ADD_U32 [[COPY]], [[COPY]], implicit-def $scc
16    ; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY [[COPY]]
17    ; GCN-NEXT: S_ENDPGM 0, implicit [[S_ADD_U32_]], implicit [[COPY1]]
18    %0:sreg_32 = COPY $sgpr0
19    %1:sreg_32 = COPY %0
20    %2:sreg_32 = S_ADD_U32 killed %0, killed %0, implicit-def $scc
21    %3:sreg_32 = COPY %1
22    S_ENDPGM 0, implicit %2, implicit %3
23...
24
25---
26name:            fold_subreg_kill
27tracksRegLiveness: true
28body:             |
29  ; GCN-LABEL: name: fold_subreg_kill
30  ; GCN: bb.0:
31  ; GCN-NEXT:   successors: %bb.1(0x80000000)
32  ; GCN-NEXT:   liveins: $sgpr0_sgpr1
33  ; GCN-NEXT: {{  $}}
34  ; GCN-NEXT:   [[COPY:%[0-9]+]]:sgpr_64(p4) = COPY $sgpr0_sgpr1
35  ; GCN-NEXT:   [[S_LOAD_DWORDX4_IMM:%[0-9]+]]:sgpr_128 = S_LOAD_DWORDX4_IMM [[COPY]](p4), 9, 0 :: (load (s128), align 4, addrspace 4)
36  ; GCN-NEXT:   [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_LOAD_DWORDX4_IMM]].sub2_sub3
37  ; GCN-NEXT:   [[COPY2:%[0-9]+]]:sreg_64 = COPY [[S_LOAD_DWORDX4_IMM]].sub0_sub1
38  ; GCN-NEXT: {{  $}}
39  ; GCN-NEXT: bb.1:
40  ; GCN-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY [[COPY2]].sub1
41  ; GCN-NEXT:   [[COPY4:%[0-9]+]]:sreg_32 = COPY [[COPY2]].sub0
42  ; GCN-NEXT:   [[COPY5:%[0-9]+]]:sreg_32 = COPY [[COPY1]].sub1
43  ; GCN-NEXT:   [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
44  ; GCN-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE killed [[COPY3]], %subreg.sub0, killed [[COPY4]], %subreg.sub1, killed [[COPY5]], %subreg.sub2, killed [[S_MOV_B32_]], %subreg.sub3
45  ; GCN-NEXT:   [[DEF:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
46  ; GCN-NEXT:   BUFFER_STORE_DWORD_OFFSET [[DEF]], killed [[REG_SEQUENCE]], 0, 0, 0, 0, implicit $exec :: (store (s32), addrspace 1)
47  bb.0:
48    liveins: $sgpr0_sgpr1
49
50    %0:sgpr_64(p4) = COPY $sgpr0_sgpr1
51    %1:sgpr_128 = S_LOAD_DWORDX4_IMM %0:sgpr_64(p4), 9, 0 :: (load (s128), align 4, addrspace 4)
52    %2:sreg_64_xexec = COPY %1.sub0_sub1:sgpr_128
53    %3:sreg_64_xexec = COPY killed %1.sub2_sub3:sgpr_128
54    %4:sreg_64 = COPY %2:sreg_64_xexec
55    %5:sreg_32 = COPY %3.sub1:sreg_64_xexec
56
57  bb.1:
58    %6:sreg_32 = COPY %4.sub1:sreg_64
59    %7:sreg_32 = COPY %4.sub0:sreg_64
60    %8:sreg_32 = COPY %5:sreg_32
61    %9:sreg_32 = S_MOV_B32 -1
62    %10:sgpr_128 = REG_SEQUENCE killed %6:sreg_32, %subreg.sub0, killed %7:sreg_32, %subreg.sub1, killed %8:sreg_32, %subreg.sub2, killed %9:sreg_32, %subreg.sub3
63    %11:vgpr_32 = IMPLICIT_DEF
64    BUFFER_STORE_DWORD_OFFSET %11:vgpr_32, killed %10:sgpr_128, 0, 0, 0, 0, implicit $exec :: (store (s32), addrspace 1)
65...
66
67